leomist-adapter-aknds-v0.91
This model is a fine-tuned version of LeoLM/leo-mistral-hessianai-7b on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.2770
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0003
- train_batch_size: 1
- eval_batch_size: 1
- seed: 42
- distributed_type: multi-GPU
- gradient_accumulation_steps: 8
- total_train_batch_size: 8
- optimizer: Adam with betas=(0.9,0.98) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 50
- num_epochs: 3
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss |
---|---|---|---|
0.3308 | 0.27 | 50 | 0.3488 |
0.2907 | 0.54 | 100 | 0.3217 |
0.3048 | 0.81 | 150 | 0.2987 |
0.2344 | 1.08 | 200 | 0.2904 |
0.2143 | 1.35 | 250 | 0.2867 |
0.2318 | 1.62 | 300 | 0.2786 |
0.2472 | 1.89 | 350 | 0.2716 |
0.1701 | 2.16 | 400 | 0.2841 |
0.169 | 2.43 | 450 | 0.2817 |
0.1617 | 2.7 | 500 | 0.2789 |
0.1506 | 2.97 | 550 | 0.2770 |
Framework versions
- PEFT 0.7.1
- Transformers 4.37.2
- Pytorch 1.14.0a0+44dac51
- Datasets 2.14.7
- Tokenizers 0.15.2
- Downloads last month
- 2
Unable to determine this model’s pipeline type. Check the
docs
.